Update app.py

app.py

@@ -1,20 +1,36 @@
-# app.py
+import gradio as gr
+import torch
+from TTS.models.vits import VitsModel
+from transformers import AutoTokenizer
+import torchaudio
 
-from fastapi import FastAPI, Request
-from fastapi.responses import JSONResponse
-from gradio_client import Client
+# Load the model and tokenizer from Hugging Face
+model = VitsModel.from_pretrained("HusseinBashir/codad_tijaabo")
+tokenizer = AutoTokenizer.from_pretrained("HusseinBashir/codad_tijaabo")
 
-app = FastAPI()
-client = Client("HusseinBashir/Somali_tts")
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+model = model.to(device).eval()
 
-@app.post("/somali-tts/")
-async def somali_tts(request: Request):
-    data = await request.json()
-    text = data.get("text")
-    if not text:
-        return JSONResponse(content={"error": "No text provided"}, status_code=400)
-    try:
-        audio_url = client.predict(text, api_name="/predict")
-        return JSONResponse(content={"audio_url": audio_url})
-    except Exception as e:
-        return JSONResponse(content={"error": str(e)}, status_code=500)
+def tts_infer(text):
+    inputs = tokenizer(text, return_tensors="pt")
+    input_ids = inputs.input_ids.to(device)
+
+    with torch.no_grad():
+        output = model(input_ids)
+        waveform = output["waveform"]
+
+    # Save or return audio
+    sample_rate = 22050  # VITS typically uses 22.05kHz
+    torchaudio.save("output.wav", waveform.cpu(), sample_rate)
+    return "output.wav"
+
+# Create Gradio UI
+interface = gr.Interface(
+    fn=tts_infer,
+    inputs=gr.Textbox(label="Geli qoraalka aad rabto in cod laga dhigo"),
+    outputs=gr.Audio(label="Codka la sameeyey"),
+    title="Codad Tijaabo TTS",
+    description="Ku qor qoraal Soomaali ah si aad cod u maqasho.",
+)
+
+interface.launch()